Books, essays, papers, and rabbit holes that changed how I think about building, systems, and craft.
The single best technical book written about how large distributed systems fail, how data is stored on disk, and how consistency models trade off.
The foundational architecture paper of modern AI. Replacing recurrent networks with parallel multi-head self-attention unlocked scaling laws.
Clarified once and for all that physical time cannot establish causal order in distributed networks, introducing Lamport timestamps and state machines.
Distinguishes essential complexity from accidental complexity, arguing that unconstrained mutable state is the root cause of software fragility.
Deconstructs Paxos into intuitive decomposed phases: leader election, log replication, and safety invariants. Essential for distributed backends.
Presents the case for 'deep modules': interfaces that are extremely simple while concealing profound internal complexity.
Demonstrated that scaling autoregressive models produces emergent general-purpose in-context problem-solving capabilities without gradient updates.
Real-world engineering tear-downs of major open source codebases (Git, Nginx, LLVM) written by their original creators.